Nearly 800,000 people in the United States (US) each year are affected by stroke, which remains the leading cause of adult disability and 5th leading cause of death. Despite the proliferation of stroke centers nationwide, access to timely stroke care and evidence-based therapies is still poor with multiple inefficiencies, redundancies, errors, and system-based barriers that lead to care fragmentation. A stroke system of care that integrates and coordinates care between non-stroke center hospitals, primary stroke centers, and comprehensive stroke centers has been recommended since it would reduce stroke-related deaths by 20,000 annually in the US and nearly 400,000 worldwide. This proposal is significant because it addresses two important gaps in knowledge about the acute evaluation and treatment of stroke patients: prehospital screening errors and delays in inter-hospital transfer. The Enhancing Stroke Prehospital and Emergency Evaluation and Delivery (E-SPEED) study will (1) enhance existing LVO screening tools, assess whether machine learning approaches improve LVO prediction, and evaluate the impact of an Enhanced LVO Screening Tool in the prehospital setting; and (2) use learning collaboratives and simulations to design, implement, and pilot test a new inter-hospital stroke transfer protocol. We bring together systems engineers and stroke health service researchers in a novel collaboration that will produce generalizable knowledge that can be widely applied to other urban regions of the country and provide preliminary data and demonstration of novel approaches that could be evaluated in large-scale comparative effectiveness studies across multiple regions.